vmoens
@vmoens.bsky.social
Member of technical staff @periodiclabs
Open-source/open science advocate
Maintainer of torchrl / tensordict / leanrl
Former MD - Neuroscience PhD
https://github.com/vmoens
Open-source/open science advocate
Maintainer of torchrl / tensordict / leanrl
Former MD - Neuroscience PhD
https://github.com/vmoens
Pinned
vmoens
@vmoens.bsky.social
· Nov 22
One of my fav projects: LeanRL, a simple RL library that provides recipes for fast RL training using torch.compile and cudagraphs.
Using these, we got >6x speed-ups compared to the original CleanRL implementations.
github.com/pytorch-labs...
Using these, we got >6x speed-ups compared to the original CleanRL implementations.
github.com/pytorch-labs...
Happy to announce that I've joined Periodic Labs as member of technical staff. We're a mission driven startup aimed at accelerating scientific discovery using AI, with a strong focus on material science (discovery of new materials such as superconductors and such). We're hiring: periodic.com
Periodic Labs
From bits to atoms.
periodic.com
October 2, 2025 at 5:28 PM
Happy to announce that I've joined Periodic Labs as member of technical staff. We're a mission driven startup aimed at accelerating scientific discovery using AI, with a strong focus on material science (discovery of new materials such as superconductors and such). We're hiring: periodic.com
The fact that all LLM libraries don't have the same data format is as surprising as the fact that there is more than one sign language dialect
March 6, 2025 at 3:42 AM
The fact that all LLM libraries don't have the same data format is as surprising as the fact that there is more than one sign language dialect
Ray is an excellent way of testing if all your `__repr__` are coded properly (but it shouldn't be)
March 6, 2025 at 2:26 AM
Ray is an excellent way of testing if all your `__repr__` are coded properly (but it shouldn't be)
Just stumbled upon RouteRL: a multiagent RL framework to facilitate the testing and development of efficient route choice strategies
coexistence-project.github.io/RouteRL/
Looks pretty cool!
coexistence-project.github.io/RouteRL/
Looks pretty cool!
RouteRL 1.0.0 documentationContentsMenuExpandLight modeDark modeAuto light/dark, in light modeAuto light/dark, in dark mode
coexistence-project.github.io
March 3, 2025 at 3:43 PM
Just stumbled upon RouteRL: a multiagent RL framework to facilitate the testing and development of efficient route choice strategies
coexistence-project.github.io/RouteRL/
Looks pretty cool!
coexistence-project.github.io/RouteRL/
Looks pretty cool!
Reposted by vmoens
What is GGUF, Safetensors, PyTorch, ONNX?
In this blog post, let's discover common formats for storing an AI model.
huggingface.co/blog/ngxson/...
In this blog post, let's discover common formats for storing an AI model.
huggingface.co/blog/ngxson/...
Common AI Model Formats
A Blog post by Xuan-Son Nguyen on Hugging Face
huggingface.co
February 27, 2025 at 5:11 PM
What is GGUF, Safetensors, PyTorch, ONNX?
In this blog post, let's discover common formats for storing an AI model.
huggingface.co/blog/ngxson/...
In this blog post, let's discover common formats for storing an AI model.
huggingface.co/blog/ngxson/...
Today we're opensourcing MLGym, an API for AI research agents.
MLGym relies on a gym environment that wraps a docker image. Each env has a task specified as a YAML file, telling in plain english what you want your LLM to achieve
👇
MLGym relies on a gym environment that wraps a docker image. Each env has a task specified as a YAML file, telling in plain english what you want your LLM to achieve
👇
February 21, 2025 at 4:44 PM
Today we're opensourcing MLGym, an API for AI research agents.
MLGym relies on a gym environment that wraps a docker image. Each env has a task specified as a YAML file, telling in plain english what you want your LLM to achieve
👇
MLGym relies on a gym environment that wraps a docker image. Each env has a task specified as a YAML file, telling in plain english what you want your LLM to achieve
👇
A few tips I share when I talk about perf with PyTorch in eager mode (with focus on small models): 🪢
February 19, 2025 at 5:13 PM
A few tips I share when I talk about perf with PyTorch in eager mode (with focus on small models): 🪢
I stand by my point that the word "agent" should be avoided at all costs.
At least in RL, anytime I see an "Agent" class it's meant to be a "whatever doesn't fit in any other bucket in my codebase".
At least in RL, anytime I see an "Agent" class it's meant to be a "whatever doesn't fit in any other bucket in my codebase".
February 11, 2025 at 8:52 AM
I stand by my point that the word "agent" should be avoided at all costs.
At least in RL, anytime I see an "Agent" class it's meant to be a "whatever doesn't fit in any other bucket in my codebase".
At least in RL, anytime I see an "Agent" class it's meant to be a "whatever doesn't fit in any other bucket in my codebase".
Everyone's like "hey I just coded and trained a SOTA LLM in my garage last week, also wrote a blogpost about it and opensourced the repo" and the only thing I did in the meantime was fix a CI and configure a remote interpreter on a server... 😢
February 6, 2025 at 3:37 PM
Everyone's like "hey I just coded and trained a SOTA LLM in my garage last week, also wrote a blogpost about it and opensourced the repo" and the only thing I did in the meantime was fix a CI and configure a remote interpreter on a server... 😢
Clap hands if you were doing RL before it was cool
January 28, 2025 at 4:04 AM
Clap hands if you were doing RL before it was cool
You’ve never really understood PyTorch until you’ve figured out what torch.scatter exactly does.
I’ve never really understood PyTorch.
I’ve never really understood PyTorch.
January 26, 2025 at 4:24 PM
You’ve never really understood PyTorch until you’ve figured out what torch.scatter exactly does.
I’ve never really understood PyTorch.
I’ve never really understood PyTorch.
Reposted by vmoens
Introducing playground.mujoco.org
Combining MuJoCo’s rich and thriving ecosystem, massively parallel GPU-accelerated simulation, and real-world results across a diverse range of robot platforms: quadrupeds, humanoids, dexterous hands, and arms.
Get started today: pip install playground
Combining MuJoCo’s rich and thriving ecosystem, massively parallel GPU-accelerated simulation, and real-world results across a diverse range of robot platforms: quadrupeds, humanoids, dexterous hands, and arms.
Get started today: pip install playground
MuJoCo Playground
An open-source framework for GPU-accelerated robot learning and sim-to-real transfer
playground.mujoco.org
January 16, 2025 at 8:48 PM
Introducing playground.mujoco.org
Combining MuJoCo’s rich and thriving ecosystem, massively parallel GPU-accelerated simulation, and real-world results across a diverse range of robot platforms: quadrupeds, humanoids, dexterous hands, and arms.
Get started today: pip install playground
Combining MuJoCo’s rich and thriving ecosystem, massively parallel GPU-accelerated simulation, and real-world results across a diverse range of robot platforms: quadrupeds, humanoids, dexterous hands, and arms.
Get started today: pip install playground
One of the bests Mindscape episodes I've heard!
I've always been interested in the epistemology of controversial scientific concepts such as emergence or consciousness. Vaguely defined concepts (or diverging definitions) are often at the root of many pointless debates.
Kudos on clarifying things!
I've always been interested in the epistemology of controversial scientific concepts such as emergence or consciousness. Vaguely defined concepts (or diverging definitions) are often at the root of many pointless debates.
Kudos on clarifying things!
Mindscape 295 | Solo: Emergence and Layers of Reality. Finally sorting out what this emergence stuff is all about, we can all carry on now. #MindscapePodcast
www.preposterousuniverse.com/podcast/2024...
www.preposterousuniverse.com/podcast/2024...
295 | Solo: Emergence and Layers of Reality – Sean Carroll
www.preposterousuniverse.com
January 9, 2025 at 2:09 PM
One of the bests Mindscape episodes I've heard!
I've always been interested in the epistemology of controversial scientific concepts such as emergence or consciousness. Vaguely defined concepts (or diverging definitions) are often at the root of many pointless debates.
Kudos on clarifying things!
I've always been interested in the epistemology of controversial scientific concepts such as emergence or consciousness. Vaguely defined concepts (or diverging definitions) are often at the root of many pointless debates.
Kudos on clarifying things!
New Year’s resolutions:
- eat healthier
- exercise more
- no more “is all you need” papers
- eat healthier
- exercise more
- no more “is all you need” papers
December 21, 2024 at 8:07 AM
New Year’s resolutions:
- eat healthier
- exercise more
- no more “is all you need” papers
- eat healthier
- exercise more
- no more “is all you need” papers
Reposted by vmoens
Yesterday the hyped Genesis simulator released. But it's up to 10x slower than existing GPU sims, not 10-80x faster or 430,000x faster than realtime since they benchmark mostly static environments
blog post with corrected open source benchmarks & details: stoneztao.substack.com/p/the-new-hy...
blog post with corrected open source benchmarks & details: stoneztao.substack.com/p/the-new-hy...
December 20, 2024 at 11:49 PM
Yesterday the hyped Genesis simulator released. But it's up to 10x slower than existing GPU sims, not 10-80x faster or 430,000x faster than realtime since they benchmark mostly static environments
blog post with corrected open source benchmarks & details: stoneztao.substack.com/p/the-new-hy...
blog post with corrected open source benchmarks & details: stoneztao.substack.com/p/the-new-hy...
Wrong answers only:
What does this `Human-computer` sticker seen at neurips hide?
What does this `Human-computer` sticker seen at neurips hide?
December 19, 2024 at 10:53 AM
Wrong answers only:
What does this `Human-computer` sticker seen at neurips hide?
What does this `Human-computer` sticker seen at neurips hide?
Check out Motivo, a behavioral foundation model for humanoid control by FAIR.
It's a one-of-its-kind unsupervised RL project, and it comes with a demo that is SO fun to play with!
metamotivo.metademolab.com
(for the record, they use compile and cudagraphs -> github.com/facebookrese...)
It's a one-of-its-kind unsupervised RL project, and it comes with a demo that is SO fun to play with!
metamotivo.metademolab.com
(for the record, they use compile and cudagraphs -> github.com/facebookrese...)
December 14, 2024 at 12:44 AM
Check out Motivo, a behavioral foundation model for humanoid control by FAIR.
It's a one-of-its-kind unsupervised RL project, and it comes with a demo that is SO fun to play with!
metamotivo.metademolab.com
(for the record, they use compile and cudagraphs -> github.com/facebookrese...)
It's a one-of-its-kind unsupervised RL project, and it comes with a demo that is SO fun to play with!
metamotivo.metademolab.com
(for the record, they use compile and cudagraphs -> github.com/facebookrese...)
I’m 100% sure this button never does anything
December 13, 2024 at 4:18 AM
I’m 100% sure this button never does anything
PyTorch has released torchcodec yesterday, a powerful video decoding toolbox
pytorch.org/blog/torchco...
github.com/pytorch/torc...
pytorch.org/blog/torchco...
github.com/pytorch/torc...
torchcodec: Easy and Efficient Video Decoding for PyTorch
We are pleased to officially announce torchcodec, a library for decoding videos into PyTorch tensors. It is fast, accurate, and easy to use. When running PyTorch models on videos, torchcodec is our re...
pytorch.org
December 12, 2024 at 10:41 PM
PyTorch has released torchcodec yesterday, a powerful video decoding toolbox
pytorch.org/blog/torchco...
github.com/pytorch/torc...
pytorch.org/blog/torchco...
github.com/pytorch/torc...
Tomorrow with Matteo Bettini we'll be presenting BenchMARL at #NeurIPS (@neuripsconf.bsky.social) in #Vancouver
December 11, 2024 at 12:47 AM
Tomorrow with Matteo Bettini we'll be presenting BenchMARL at #NeurIPS (@neuripsconf.bsky.social) in #Vancouver
I'm excited to be in #Vancouver for #NeurIPS 2024 where I brought a bunch of @LEGO_Group bricks in my bag!
*no, I'm not Santa!
*no, I'm not Santa!
December 9, 2024 at 7:19 PM
I'm excited to be in #Vancouver for #NeurIPS 2024 where I brought a bunch of @LEGO_Group bricks in my bag!
*no, I'm not Santa!
*no, I'm not Santa!
We’re looking for an intern (research scientist/PhD) to join the PyTorch team in NYC this summer and work on GPU kernel generation, more info here:
www.metacareers.com/jobs/6044463...
You’ll be working with @soumithchintala.bsky.social, @marksaroufim.bsky.social and myself
www.metacareers.com/jobs/6044463...
You’ll be working with @soumithchintala.bsky.social, @marksaroufim.bsky.social and myself
Research Scientist Intern, PyTorch Core (PhD)
Meta's mission is to build the future of human connection and the technology that makes it possible.
www.metacareers.com
December 7, 2024 at 2:50 PM
We’re looking for an intern (research scientist/PhD) to join the PyTorch team in NYC this summer and work on GPU kernel generation, more info here:
www.metacareers.com/jobs/6044463...
You’ll be working with @soumithchintala.bsky.social, @marksaroufim.bsky.social and myself
www.metacareers.com/jobs/6044463...
You’ll be working with @soumithchintala.bsky.social, @marksaroufim.bsky.social and myself
There are two kinds of developers:
- Those who know what they would have done differently had they known what a pain it was going to be
- Those who change things accordingly and ruin everyone's life forever
- Those who know what they would have done differently had they known what a pain it was going to be
- Those who change things accordingly and ruin everyone's life forever
December 6, 2024 at 12:09 PM
There are two kinds of developers:
- Those who know what they would have done differently had they known what a pain it was going to be
- Those who change things accordingly and ruin everyone's life forever
- Those who know what they would have done differently had they known what a pain it was going to be
- Those who change things accordingly and ruin everyone's life forever